Abstract

   Background/Objectives: This study explores the allelopathic effects of
   volatile organic compounds (VOCs) emitted by the invasive species
   Ailanthus altissima (Mill.) Swingle on the seed germination of Bidens
   pilosa. A. altissima is known for releasing allelopathic VOCs that
   suppress the growth of neighbouring plants, contributing to its
   invasive potential. Methods: To examine these effects, we exposed B.
   pilosa seeds to varying concentrations of A. altissima VOCs, assessing
   germination rates and metabolic changes through untargeted
   metabolomics. Results: Our findings revealed that VOCs from A.
   altissima significantly inhibited the germination speed and overall
   germination rates of B. pilosa in a dose-dependent manner. Metabolomic
   profiling showed disruptions in energy and amino acid metabolism
   pathways, specifically involving delayed breakdown of starch and key
   metabolites, indicating inhibition of critical metabolic processes
   during early germination stages. This metabolic delay likely impairs B.
   pilosa’s establishment and competitiveness, enhancing A. altissima’s
   ecological dominance. Conclusions: The results underscore the potential
   of VOC-based allelopathy as a mechanism of plant invasion, offering
   insights into the role of VOCs in interspecies plant competition and
   ecosystem dynamics.

   Keywords: allelopathy, alien species, metabolomics, plant metabolism,
   phytotoxicity

1. Introduction

   Plant–plant communication through volatile organic compounds (VOCs)
   represents a fascinating area of research that elucidates how plants
   can influence one another’s growth and defence mechanisms [[28]1]. This
   communication often occurs when neighbouring plants release VOCs in
   response to herbivore attacks or other stressors, which can prime
   adjacent plants to enhance their defensive strategies [[29]2].

   Plant volatiles are chemicals that give plants a characteristic and
   typical odour, flavour, and several other properties. They are complex
   mixtures mainly composed of terpenes, synthesised within plants as
   secondary metabolites, providing plants with many ecological advantages
   [[30]3]. They play a pivotal role in plant reproduction, acting as
   pollinator attractants, providing protection against herbivores, and
   mediating plant communication, being determinants for vegetation
   patterning [[31]3,[32]4]. These chemicals are thought to be important
   allelopathic agents, especially in arid and semi-arid conditions where
   they act in the vapour phase [[33]5,[34]6,[35]7], and early studies
   demonstrated that plant volatiles are potent seed germination
   inhibitors, playing an important role in ecosystems by reducing
   seedling establishment ability and growth [[36]6,[37]8,[38]9]. In
   addition, several studies demonstrated that specific volatile compounds
   can significantly impede seed germination rates. For instance, Kang et
   al. [[39]10] identified L-Fenchone and 1,8-Cineole from the leaves of
   Star Anise (Illicium verum), which exhibited notable allelopathic
   activity when tested on lettuce seeds using the Dish pack method. Their
   findings indicated that these volatiles could effectively inhibit
   germination, highlighting the potential of using such compounds to
   manage weed populations [[40]10]. Similarly, Mishyna et al. [[41]11]
   reported that octanal, a compound released from Heracleum sosnowskyi
   fruit, strongly inhibited both seed germination and radicle elongation
   in lettuce, suggesting that octanal is a key contributor to the
   allelopathic effects of this species [[42]11].

   The mechanism by which these volatiles exert their inhibitory effects
   often involves interference with physiological processes in the target
   plants. Maffei et al. [[43]12] noted that certain monoterpenes, such as
   carvacrol and D-limonene, can disrupt the cytochromic respiration
   pathway and inhibit growth and seed germination in neighbouring plants.
   Also, the delay of seed germination and growth is an important
   ecological effect, which plays a significant ecological role in natural
   ecosystems, increasing the competitive ability of the species, which
   can retard the growth of the competitors. McCalla and Norstadt [[44]13]
   defined germination as the most sensitive vegetative stage to
   phytotoxins. A short period of inhibition or stimulation at this stage
   could strongly increase or reduce the ability to compete with other
   plants [[45]13,[46]14].

   One worldwide-spread species known for its strong allelopathic
   properties is Ailanthus altissima (Mill.) Swingle, the “tree of
   heaven”. Originally from Asia, A. altissima has become a widespread
   invasive species in many parts of the world, often dominating native
   ecosystems [[47]15]. Its allelopathic effects have been well
   documented, with studies revealing its ability to inhibit the growth
   and germination of various plant species, thereby contributing to its
   competitive advantage in non-native regions [[48]16,[49]17]. The VOCs
   emitted by A. altissima could be of particular interest, as they may
   play a role in its invasive success by inhibiting or altering the
   growth patterns of neighbouring plants. While A. altissima’s general
   allelopathic properties are recognised, there is limited research on
   the specific effects of its VOCs on the germination process of
   neighbouring species. This represents a notable gap in the study of
   plant–plant chemical interactions. Given the potential role of VOCs in
   delaying or inhibiting germination, research into their specific impact
   could contribute valuable insights into the invasiveness ability of A.
   altissima.

   One neighbouring species of interest is Bidens pilosa L. (Asteraceae),
   commonly known as blackjack. Originally from Central and South America,
   this fast-growing weed is now widely distributed across tropical and
   subtropical regions, posing a threat to vital crops [[50]18]. Several
   traits contribute to the invasiveness of B. pilosa. It exhibits a high
   growth rate, capable of producing multiple generations annually, and
   generates an abundant number of seeds. These seeds readily attach to
   animal fur, bird feathers, and human clothing, facilitating its spread
   to new areas. The seeds germinate rapidly under diverse conditions,
   enhancing the plant’s adaptability. B. pilosa thrives in a wide range
   of environmental settings [[51]19].

   Renowned for its aggressive colonisation, prolific seed production, and
   adaptability to disturbed habitats, B. pilosa presents a significant
   challenge to agriculture and native ecosystems. Its allelopathic
   potential further strengthens its competitive edge, making it an
   excellent model for studying plant–plant interactions [[52]20].
   Investigating the effects of Ailanthus altissima’s volatile organic
   compounds (VOCs) on this invasive weed could provide valuable insights
   into the ecological impacts of allelopathic interactions and their role
   in the success of invasive species.

   The present study aims to investigate the allelopathic potential of
   VOCs emitted by A. altissima on the germination of B. pilosa. By
   examining how these VOCs influence the metabolic activity and
   germination timing of B. pilosa seeds, this research seeks to
   demonstrate the potential of VOC-based allelopathy as a natural means
   of plant invasion. Our findings reveal that the VOCs released by A.
   altissima slow down the germination metabolism of B. pilosa, delaying
   its germination and potentially reducing seedlings’ establishment and
   competitiveness.

2. Materials and Methods

2.1. Plant Material

   Aerial parts of Ailanthus altissima were sampled in June in Calabria
   (Southern Italy). The plant material was collected in an invaded
   ecosystem close to Reggio Calabria (Italy) (latitude N 38°16′9030″,
   longitude E 15°67′5358″), and its collection did not require any
   specific permission.

2.2. Germination Indexes

   To allow for the synchronisation of seed germination, the seeds of
   Bidens Pilosa were poured into a beaker filled with deionised water and
   stored for two days in a refrigerator at 4 °C. The seeds were then
   successively blotted on filter paper and dried at room temperature
   under a chemical cabin with laminar flow. Once dried, the seeds were
   sterilised for 15 min in 5% NaClO (Merk Life Science, Milan, Italy)
   solution, and 20 seeds (for each replicate and concentration) were sown
   in Petri dishes (9 cm in Ø), with the bottom covered by a double layer
   of filter paper and 5 mL of deionised water.

   To verify if the VOCs released by the leaves of A. altissima could
   inhibit seed germination, the Petri dishes (without the lids) were
   exposed for 8 days in an open box to different amounts (0, 50, 100,
   150, 200, and 250 g) of freshly collected A. altissima leaves (changed
   daily) following the protocol previously described by Araniti et al.
   [[53]6].

   The germination was monitored daily, and then the germination index
   [G[T](%)] and speed of germination (S) were calculated using the
   following formula:
   G[T](%) = [(NT × 100)/N]

   where N[T] is the number of germinated seeds and N is the number of
   seeds used in the bioassay.
   S = (N[1] × 1) + (N[2] − N[1]) × 1/2

   N[1], N[2], N[3], …… N[n−1], N[n]; the proportion of the germinated
   seed obtained at the first (1), the second (2), the third (3), (n − 1),
   n hours or days after sowing.

   This experiment allowed for us to determine the key concentration of
   150 g of fresh material, which was then used for the metabolomic
   analysis.

2.3. VOC Characterisation: HS-SPME-GC–MS Analysis

   VOC characterisation was carried out through an HS-SPME-GC–MS analysis
   using a gas chromatographer (Agilent 7890A GC—Agilent Scientific
   Instruments, Santa Clara, CA, USA) coupled with a single quadrupole
   mass spectrometer (Agilent 5975C—Agilent Scientific Instruments, Santa
   Clara, CA, USA) and equipped with a MEGA-5MS capillary column (30 m ×
   0.25 mm × 0.25 µm plus 10 m pre-column) (MEGA S.r.l., Legnano (MI),
   Italy). Plant material (1 g per replicate) was incubated for 20 min in
   a 20 mL glass vial at room temperature. Subsequently, an SPME grey
   fibre (StableFlex, coated with divinylbenzene/Carboxen on
   polydimethylsiloxane; 50/30 μm coating; Merk Life Science, Milan,
   Italy) was exposed to the volatile organic compounds (VOCs) released by
   the plants for 20 min to allow adsorption onto the fibre. The adsorbed
   sample was then desorbed into the GC–MS, operating in splitless mode.
   The injector was settled at 200 °C, and the transfer line was set at
   250 °C, whereas the source and the quadrupole were set at 280 °C and
   150 °C, respectively. The temperature program was set as follows: an
   isocratic hold at 45 °C for 7 min, followed by a temperature increase
   from 45 °C to 80 °C at a rate of 10 °C per minute, then from 80 °C to
   200 °C at a rate of 20 °C per minute, and a final isocratic hold at 200
   °C for 3 min.

2.4. Untargeted Metabolomic Analysis

   To evaluate the effects of coumarin on B. pilosa seed metabolism during
   the germination process, seeds were treated as previously described
   using 150 g of A. altissima leaves. This concentration was chosen since
   the higher concentration did not reduce total germination but
   significantly reduced its speed.

   Seeds were collected at different time points (T0 = day 0; T1 = day 2;
   T2 = day 4; T3 = day 6; T4 = day 8), snap-frozen in liquid nitrogen,
   and powdered. A total of 100 mg of plant material for each treatment (0
   g and 150 g) replicate was transferred into a 2 mL vial.

   Extraction and sample derivatisation were carried out following the
   protocol proposed by Lisec et al. [[54]21] and modified by Araniti et
   al. [[55]22].

   For extraction, 1400 μL of chilled methanol (−20 °C) was added along
   with 60 μL of ribitol (0.2 mg/mL in water) as an internal standard for
   the polar phase, followed by vortexing for 10 s. The tubes were heated
   in a thermomixer at 70 °C with shaking at 950 rpm for 10 min and then
   centrifuged at 11,000 g for 10 min. The resulting supernatants were
   transferred to glass vials, where 750 μL of chilled chloroform (−20 °C)
   and 1500 μL of cold water (4 °C) were added sequentially. After
   vortexing for 10 s, the samples were centrifuged again at 2200 g for 15
   min.

   The polar phase (150 μL) from each sample was collected into 1.5 mL
   tubes and dried using a vacuum concentrator without heating.

   For derivatisation, 40 μL of methoxyamine hydrochloride (20 mg/mL in
   pyridine) (Merk Life Science, Milan, Italy) was added to the dried
   samples, which were incubated in a thermomixer at 37 °C and shaken at
   950 rpm for 2 h. This step was followed by silylation, achieved by
   adding 70 μL of MSTFA (Merk Life Science, Milan, Italy) to each sample
   and shaking at 37 °C for 30 min. Finally, 110 μL of the derivatised
   samples was transferred to GC/MS-compatible glass vials for analysis.

   The derivatised extracts were analysed using a MEGA-5MS capillary
   column (30 m × 0.25 mm × 0.25 µm, equipped with a 10 m pre-column) in a
   gas chromatography system (Agilent 7890A GC) coupled with a single
   quadrupole mass spectrometer (Agilent 5975C). The injector and ion
   source temperatures were set to 250 °C and 260 °C, respectively. A 1 µL
   aliquot of each sample was injected in splitless mode, with helium as
   the carrier gas at a 1 mL/min flow rate. The temperature program
   included an initial isothermal period at 70 °C for 5 min, followed by a
   temperature increase of 5 °C/min to 350 °C, and a final hold at 330 °C
   for 5 min. Mass spectra were acquired in electron ionisation (EI) mode
   at 70 eV, scanning from 40 to 600 m/z with a scan time of 0.2 s. A
   solvent delay of 9 min was applied. Samples, n-alkane standards, and
   blank pyridine solvent were periodically injected to assess instrument
   performance, confirm tentative identifications, and track any retention
   index (RI) shifts. Solvent blanks were run between samples to monitor
   potential contamination, and each mass was verified against the blanks
   to rule out contamination sources.

2.5. GC/MS Data Analysis Using MS-DIAL

   Raw peak extraction in this study was conducted using MS-DIAL with an
   open-source, publicly available EI spectral library. The baseline of
   the data was filtered and calibrated, and subsequent steps—such as peak
   alignment, deconvolution analysis, peak identification, and peak height
   integration—were performed following the methodology outlined in Misra
   et al. [[56]23]. Detection parameters included an average peak width of
   20 scans, a minimum peak height threshold of 1000 amplitudes, and a
   sigma window value of 0.5, with an EI spectra cut-off set at 5000
   amplitudes for deconvolution. The identification of peaks utilised a
   retention time tolerance of 0.2 min, an m/z tolerance of 0.5 Da, an EI
   similarity threshold of 60%, and a minimum identification score of 80%.
   Alignment parameters were adjusted with a retention time tolerance of
   0.5 min and a retention time factor of 0.5.

   For MS-DIAL data annotations, compound identification was achieved
   using a homemade library built following the protocol proposed by Misra
   [[57]24].

   Only confidently annotated and quantified metabolites were reported.
   Metabolite identification followed the Metabolomics Standards
   Initiative (MSI) guidelines [[58]25]: Level 2 identification was based
   on spectral database matching with a match factor above 80%; Level 3
   allowed for compound group identification through specific ions and
   retention time regions.

   Concerning VOC data, all the putatively annotated metabolites resulting
   from MS-DIAL to have a total score higher than 80% were checked
   one-by-one on several open-source plant databases
   ([59]www.plantcyc.org, accessed on 1 October 2024) and/or by checking
   in the literature if these molecules were previously annotated on the
   same genus or species. Those metabolites not belonging to the plant
   metabolome were considered unknowns and removed from the analysis.

2.6. Statistical Analysis

   The experiments were conducted using a randomised design with five
   replications for the germination tests and four replications for
   metabolomics experiments. Germination parameters were previously tested
   for normality and homogeneity of the variance and successively analysed
   through one-way ANOVA using the LSD test as post hoc (p ≤ 0.05).

   Metabolomic data were analysed using Metaboanalyst 6.0 [[60]26]. Using
   the R-based open-source software MetaboAnalyst 6.0, data were
   normalised using the internal standard (ribitol—0.02 mg/mL) to account
   for variations in sample preparation, instrument performance, and
   extraction efficiency, ensuring consistency and comparability across
   samples, and the missing values were replaced with half of the minimum
   value in the data set. Successively, data were Log[10] transformed to
   reduce the impact of large differences in metabolite concentrations,
   making the data more manageable and highlighting relative changes
   rather than absolute differences. Finally, transformed data were Pareto
   scaled (mean-centred and divided by the square root of the standard
   deviation of each variable) to further enhance this by reducing the
   dominance of highly abundant metabolites while still preserving the
   data’s structure, making subtle variations more apparent and improving
   the interpretability of multivariate analyses.

   Data were then classified through principal component analysis (PCA) to
   provide an overview of the quality of the data acquisition step and
   evaluate group separation. Furthermore, cluster analysis was performed
   using the Euclidean distance measure and the Ward algorithm for cluster
   formation to examine the classification further.

   A within-subjects two-way ANOVA was used, and the significance
   threshold was defined as the corrected p-value ≤ 0.05. The false
   discovery rate was chosen for multiple testing corrections. Moreover,
   GC–MS data were analysed using Multivariate Empirical Bayes Analysis of
   Variance (MEBA) for Time Series, and features with significant changes
   over time (using the Hotelling-T2 values) were examined for their
   dose–response during the germination process.

3. Results

3.1. Germination Index

   The results revealed a significant inhibition potential of A. altisisma
   VOCs on B. pilosa seed germination. In particular, the chemicals
   released from the plant material significantly reduced total
   germination (G[T] %) and speed (S %). The S parameter was significantly
   inhibited (≈12%), even at the lower treatments (50 g and 100 g),
   increasing up to 73% at the highest concentration assayed (250 g of
   plant material) ([61]Figure 1). Conversely, the lowest treatments (from
   50 g to 150g) did not affect the GT parameter. The 200 g treatment
   caused 27% inhibition, reaching 60% at the highest concentration
   ([62]Figure 1). It should be noted that, enlarging the end of the
   experiments to 13 days, almost all seeds were germinating, even at 250
   g treatment.

Figure 1.

   [63]Figure 1
   [64]Open in a new tab

   Dose-response curves of the (a) germination index [GT(%)] and (b)
   germination speed (S) of B. pilosa seeds exposed for 8 days to
   Ailanthus altissima VOCs. Data are expressed as mean ± SD and analysed
   through one-way ANOVA using the LSD test as post hoc (p ≤ 0.05).
   Different letters along the curves (a–e) indicate statistical
   differences among the treatments with p ≤ 0.05. N = 4.

3.2. VOC Characterisation

   [65]Table 1 presents a chemical breakdown of the volatile organic
   compounds (VOCs) derived from Ailanthus altissima. The major chemical
   classes identified include alcohols, polyols, carbonyl compounds, fatty
   acid esters, alkanes, and monoterpenoids ([66]Table 1).

Table 1.

   Chemical characterisation of Ailanthus altissima volatile organic
   compounds.
   RT RI Metabolite Name SubClass RAP%
   2.517 681.83 2-ethyl-furan Furanoid 0.45 (±0.08)
   2.867 706.23 2-Methyl-1-butanol Alcohols and polyols 0.57 (±0.17)
   2.99 714.66 (E)-2-methyl-2-Butenal Carbonyl compounds 0.19 (±0.10)
   3.446 746.61 1-Pentanol Alcohols and polyols 0.12 (±0.04)
   3.594 756.92 1-Penten-3-ol, 2-methyl- Alcohols and polyols 0.41 (±0.06)
   3.684 763.25 Methyl α-methylbutanoate Fatty acid esters 0.11 (±0.04)
   4.192 798.74 3-Hexenal Carbonyl compounds 1.11 (±1.15)
   4.23 800.16 Heptane, 2,4-dimethyl- Alkanes 0.77 (±0.04)
   4.244 800.81 Hexanal Carbonyl compounds 0.57 (±0.28)
   4.435 805.44 3-Octene, (Z)- Unsaturaed aliphatic hydrocarbons 0.28
   (±0.27)
   5.332 826.58 3-Hexene, 1-methoxy-, (Z)- Ethers 0.24 (±0.05)
   6.18 846.69 Butanoic acid, 2-methyl-, ethyl ester Fatty acid esters
   0.30 (±0.23)
   6.224 847.76 3-Hexen-1-ol, (E)- Fatty alcohols 0.60 (±0.22)
   6.395 851.78 3-Hexen-1-ol, (Z)- Fatty alcohols 47.52 (±12.25)
   6.573 856.22 4-Hexen-1-ol, (Z)- Fatty alcohols 10.73 (±1.72)
   6.976 865.51 2-Hexen-1-ol, (Z)- Fatty alcohols 1.52 (±1.07)
   7.63 881.05 1-Butanol, 2-methyl-, acetate Carboxylic acids derivatives
   0.38 (±0.17)
   8.423 899.9 4,4-Dimethyl octane Alkanes 0.12 (±0.05)
   9.052 921 Amyl acetate Carboxylic acids derivatives 0.11 (±0.04)
   9.349 931.05 Dimethylallyl acetate Carboxylic acids derivatives 0.32
   (±0.26)
   9.541 937.55 alpha-Pinene Monoterpenoids 1.49 (±0.22)
   10.74 978.03 β-Thujene Monoterpenoids 0.13 (±0.01)
   11.205 993.75 p-Mentha-1(7),8-diene Monoterpenoids 1.66 (±1.26)
   11.54 1009.79 3-Methyl-4-penten-1-ol acetate Carboxylic acids
   derivatives 58.58 (±47.89)
   11.659 1017.68 Hexyl acetate Carboxylic acids derivatives 4.33 (±1.23)
   11.711 1020.99 2-Hexen-1-yl-acetate Carboxylic acids derivatives 3.50
   (±2.65)
   11.872 1032.82 3-Carene Monoterpenoids 0.31 (±0.20)
   11.892 1032.82 L-Limonene Monoterpenoids 0.76 (±0.31)
   12.064 1044.02 1,4-p-Menthadien-7-ol Alcohols and polyols 0.28 (±0.03)
   12.083 1045.27 o-Cresol Cresols 0.18 (±0.16)
   12.23 1054.92 cis-β-Ocimene Monoterpenoids 1.03 (±0.58)
   12.345 1062.39 Artemisia alcohol Alcohols and polyols 0.19 (±0.20)
   12.445 1068.61 Artemesia ketone Carbonyl compounds 0.55 (±0.12)
   12.792 1091.65 Isoterpinolene Monoterpenoids 0.49 (±0.20)
   12.935 1101.41 Linalyl formate Monoterpenoids 0.39 (±0.26)
   12.987 1106.3 Nonanal Carbonyl compounds 0.21 (±0.09)
   13.134 1119.96 Perillene Monoterpenoids 8.25 (±5.77)
   13.43 1147.7 (Z)-3-Hexenyl butanoate Fatty acid esters 0.24 (±0.10)
   13.678 1170.83 α-Acetoxytoluene Benzyloxycarbonyls 0.20 (±0.23)
   13.768 1179.3 Ethyl benzoate Benzoic acids and derivatives 0.15 (±0.03)
   13.864 1188.19 (E)-2,6-Dimethyl-3,7-octadien-2,6-diol Alcohols and
   polyols 0.15 (±0.13)
   14.022 1203.86 Methyl salicylate Benzoic acids and derivatives 7.92
   (±8.17)
   14.284 1235.04 cis-3-Hexenyl valerate Fatty acid esters 2.84 (±1.02)
   14.319 1239.2 Hexyl 2-methylbutyrate Fatty acid esters 0.19 (±0.02)
   14.797 1296.07 2-Undecanone Carbonyl compounds 0.04 (±0.05)
   14.84 1301.33 Dihydroedulan II Sesquiterpenoids 0.05 (±0.01)
   15.192 1348.96 α-Longipinene Sesquiterpenoids 0.67 (±0.60)
   15.297 1363.12 Copaene Sesquiterpenoids 11.52 (±9.81)
   15.353 1370.63 Longicyclene Sesquiterpenoids 0.48 (±0.35)
   [67]Open in a new tab

   RT: retention time; RI: retention index; RAP%: area percentage of a
   specific metabolite calculated on the total area of the metabolites
   identified. The information concerning the peak area, spectra total
   score similarity (% of match with the library), retention index (RI)
   similarity, S/N ratio, and EI spectrum are reported in
   [68]Supplementary Table S1.

   The Relative Area Percentage (RAP%) column quantifies the abundance of
   each compound within the sample, offering insights into which chemicals
   dominate the volatile profile of Ailanthus altissima. The most abundant
   compound identified is (Z)-3-hexen-1-ol, a fatty alcohol comprising
   11.09% of the total volatile composition ([69]Table 1). This compound,
   alongside 3-hexenal (10.18%) and hexyl trifluoroacetate (9.46%),
   suggests a volatile profile rich in fatty alcohols and esters
   ([70]Table 1). Alcohols and polyols form a significant portion of the
   volatile makeup, with notable entries like 2-methyl-1-butanol (8.33%)
   and 1-pentanol (5%) ([71]Table 1). Carbonyl compounds, such as
   3-hexenal and hexanal, are similarly well represented. Meanwhile,
   unsaturated aliphatic hydrocarbons (e.g., (Z)-3-octene, at 8.88%) and
   monoterpenoids (compounds like β-thujene and α-pinene, albeit at lower
   concentrations) add another layer to the VOC profile ([72]Table 1). In
   summary, the chemical characterisation of A. altissima reveals a rich
   blend of volatile compounds, dominated by fatty alcohols, carbonyl
   compounds, and alcohols.

3.3. GC–MS Untargeted Metabolomics

   A GC/MS-driven untargeted metabolic analysis was conducted to gain more
   insights into the A. altissima VOC-mediated metabolic changes induced
   in B. pilosa seed metabolism during germination.

   The GC–MS-driven analysis was performed on seeds exposed to 150 g of
   freshly cut leaves of A. altissima for several days (T0 = day 0; T1 =
   day 2; T2 = day 4; T3 = day 6; T4 = day 8). The analysis revealed
   grouped and individual metabolites that allowed for sample
   discrimination. Among all the analysed stages, the metabolomic analysis
   allowed for us to annotate and quantify 61 metabolites, mainly
   belonging to the classes of amino acids, organic acids, sugars, and
   sugar alcohols, among others ([73]Supplementary Table S1).

   After sample normalisation, transformation, and scaling, the data were
   statistically analysed through unsupervised principal component
   analysis (PCA) to obtain a global view of the kinetic metabolic
   patterns of developing B. pilosa seeds during VOC exposition. The PCA
   was built by virtue of the first two components (PC1 vs. PC2), which
   accounted for 61.5% of the total variance ([74]Figure 2).

Figure 2.

   [75]Figure 2
   [76]Open in a new tab

   Effects of A. altissima VOCs on Bidens pilosa seed metabolome. (a)
   Unsupervised PCA scores plot between the two selected PCs (the
   explained variances are shown in brackets); (b) Clustering result shown
   as a dendrogram (distance measure using Euclidean, and clustering
   algorithm using ward). N = 4.

   The PCA analysis provides a detailed perspective on the time-dependent
   effects of Ailanthus altissima volatiles (VOCs) on seeds, revealing
   distinct patterns of variation between the treated and untreated groups
   ([77]Figure 2). The analysis highlights the role of PC1 in capturing
   temporal variation, while PC2 primarily differentiates seeds based on
   the treatment, i.e., the presence or absence of VOC exposure
   ([78]Figure 2).

   The progression along PC1, which explains 47.4% of the total variance,
   is largely driven by the time-related dynamics in both treated and
   untreated seeds. For untreated seeds (C-T0 through C-T4), there is a
   clear temporal progression from left to right along PC1. This trend
   suggests that, even in the absence of VOCs, the seeds undergo gradual
   physiological changes over time, likely reflecting natural
   developmental processes. Similarly, treated seeds (T-T1 through T-T4)
   show a temporal trend along PC1 but with a much more pronounced shift,
   indicating that exposure to VOCs accelerates or amplifies
   time-dependent changes in seed physiology ([79]Figure 2).

   In contrast, PC2, which accounts for 14.1% of the variance, serves as
   the axis of differentiation between the treatment groups. Untreated
   seeds cluster together with minimal dispersion along PC2, reflecting
   the uniformity of their responses in the absence of VOCs. Treated
   seeds, however, are clearly separated from untreated seeds along this
   axis, demonstrating the substantial impact of VOC exposure. This
   separation is consistent across all time points, indicating that VOCs
   introduce a distinct set of changes in seed responses that are
   independent of natural temporal variation ([80]Figure 2).

   Interestingly, the combined interpretation of PC1 and PC2 reveals that
   the effects of VOCs become increasingly pronounced over time. At the
   initial stages (T-T1), the treated seeds are still relatively close to
   the untreated groups along PC1, indicating that time-related changes
   are just beginning to manifest. However, as time progresses, the
   treated seeds move further along PC1, with significant dispersion along
   PC2 as well. This suggests that, while PC1 captures the cumulative,
   time-dependent effects of VOCs, PC2 highlights the specific impact of
   the treatment itself, which becomes more variable and pronounced with
   longer exposure ([81]Figure 2).

   By T-T4, the treated seeds form a distinct and isolated cluster, far
   removed from the untreated seeds on both axes. This reinforces the idea
   that the VOCs not only accelerate time-dependent changes (as seen along
   PC1) but also introduce unique responses that are absent in the
   untreated seeds (as captured by PC2). The lack of overlap between the
   treated and untreated clusters at all time points highlights the
   robustness of the VOC effect ([82]Figure 2).

   The evaluation of the PCA loading plots pointed out that PC1 was
   positively influenced by glutamine, sorbose, ribose, fructose,
   isoleucine, serine, and proline, whereas melezitose and raffinose,
   methylmalonic acid, and galactinol negatively influenced it. On the
   contrary, PC2 was positively dominated by proline, maltose, putrescine,
   4-hydroxybutiric acid, and glucose, whereas 4-aminobutyric acid,
   sorbose, glyceric acid, succinic acid, asparagine, and norleucine,
   among others, negatively affected it ([83]Supplementary Table S1).

   The separation trend observed during the unsupervised PCA was further
   investigated by the cluster analysis, which highlighted the formation
   of three main groups ([84]Figure 2). The first group was formed by the
   control seeds at T0 (C-T0) and the treated seeds at T1 and T2 (T-T1 and
   T-T2). The second group was formed only by the controls at T1 and T2
   (C-T1 and C-T2), whereas the remaining treatments formed the third
   group (C-T3, C-T4, T-T3, and T-T4) ([85]Figure 2).

   A KEGG-based enrichment analysis, performed to identify the classes of
   metabolites that were overrepresented in a large set of metabolites,
   revealed enrichment of glycerolipid and propanoate metabolism;
   gluconeogenesis; sphingolipid metabolism; mitochondrial electron
   transport chain; valine, leucine, and isoleucine degradation; arginine
   and proline metabolism; citric acid cycle; and glycolysis, among others
   ([86]Figure 3 and [87]Supplementary Table S1).

Figure 3.

   [88]Figure 3
   [89]Open in a new tab

   Pathway enrichment analysis revealed different metabolic pathways
   enriched during Bidens pilosa germination in response to different
   coumarin doses (0–800 µM). (p-value cut off ≤ 0.05).

   Successively, data were analysed through pathway analysis, which
   combines enrichment and topology analysis. It pointed out that 30
   pathways (14 with an impact higher > 0.2) were significantly changed
   during seed germination, and those more affected were involved in amino
   acid and sugar metabolism ([90]Table 2 and [91]Supplementary Table S1).

Table 2.

   Result from “Pathway Analysis” (topology + enrichment analysis) carried
   out on the metabolite identified in B. pilosa seeds during the
   germination process (T0–T4) in response to the VOCs produced by A.
   altissima (0 g and 150 g).
   Pathway Total Cmpd Hits Raw p FDR Impact
   Alanine aspartate and glutamate metabolism 22 7 3.37 × 10^−11 9.26 ×
   10^−10 0.77698
   Galactose metabolism 27 9 2.02 × 10^−10 2.77 × 10^−9 0.64052
   Starch and sucrose metabolism 22 7 3.83 × 10^−12 2.11 × 10^−10 0.60856
   Isoquinoline alkaloid biosynthesis 6 1 4.33 × 10^−6 4.98 × 10^−6 0.5
   Phenylalanine metabolism 12 1 2.16 × 10^−6 3.49 × 10^−6 0.42308
   Arginine biosynthesis 18 6 8.27 × 10^−9 4.13 × 10^−8 0.36117
   Glycine serine and threonine metabolism 33 4 4.92 × 10^−9 3.01 × 10^−8
   0.32415
   Fructose and mannose metabolism 18 2 1.12 × 10^−6 2.00 × 10^−6 0.29099
   Arginine and proline metabolism 32 5 7.14 × 10^−9 3.93 × 10^−8 0.2634
   Beta-alanine metabolism 18 2 1.04 × 10^−8 4.42 × 10^−8 0.25397
   Citrate cycle (TCA cycle) 20 4 1.94 × 10^−10 2.77 × 10^−9 0.21839
   Amino sugar and nucleotide sugar metabolism 52 3 2.36 × 10^−7 6.50 ×
   10^−7 0.2096
   Tryptophan metabolism 29 1 2.54 × 10^−7 6.64 × 10^−7 0.20611
   Tyrosine metabolism 17 2 8.72 × 10^−7 1.81 × 10^−6 0.20112
   [92]Open in a new tab

   Total Cmpd: the total number of compounds in the pathway; Hits: the
   matched number from the uploaded data; Raw p: the original p-value;
   FDR: the false discovery rate applied to the nominal p-values to
   control for false-positive findings; Impact: the pathway impact value
   calculated from pathway topology analysis.

   A two-way ANOVA was then conducted to determine which factors
   (developmental time, VOC treatment, and their interaction) cause the
   variation in each metabolite. Among the identified 60 metabolites, 54
   were influenced by the VOCs (all excluded glucose-1-phosphate,
   trisaccharide, 4-aminobutyric acid, sinigrin, salicyl
   alcohol-b-glucoside, stearic acid, melezitose), all were influenced by
   the time, and 58 were influenced by the interaction between time and
   treatment (all metabolites excluded norleucine and 1,6-anhydroglucose)
   ([93]Figure 4 and [94]Supplementary Table S1).

Figure 4.

   [95]Figure 4
   [96]Open in a new tab

   Venn diagram summary of results from two-way ANOVA.

   The multifactorial two-way ANOVA confirmed the significance of the
   data, which were further analysed through Multivariate Empirical Bayes
   Analysis of Variance (MEBA) for Time Series, a statistical approach
   commonly used to identify and analyse temporal patterns across multiple
   variables within time series data. The approach is designed to compare
   the time-course profiles under different conditions. The result is a
   list of variables ranked by their differences in temporal profiles
   across different biological conditions. The Hotelling-T2 was used to
   rank the variables with different temporal profiles between the
   biological conditions under study.

   The results were grouped into chemical classes (amino acids, [97]Figure
   5; organic acids and polyamine, [98]Figure 6; sugars, [99]Figure 7;
   miscellanea, [100]Figure 8) and graphically represented ([101]Figure 5,
   [102]Figure 6, [103]Figure 7 and [104]Figure 8). The complete list of
   the metabolites and the Hotelling-T2 values for each metabolite are
   reported in [105]Supplementary Table S1.

Figure 5.

   [106]Figure 5
   [107]Open in a new tab

   Amino acids with distinctive temporal profiles between control plants
   (WT) and plants exposed to A. altissima VOCs (TR). The temporal
   profiles are presented as line graphs based on the two conditions under
   study: green for the WT and red for the TR. The graphs represent the
   normalised relative abundance profiles of the amino acids with a good
   classification score and previously significantly affected by the
   two-way ANOVA (p ≤ 0.05). N = 4.

Figure 6.

   [108]Figure 6
   [109]Open in a new tab

   Organic acids and polyamines with distinctive temporal profiles between
   control plants (WT) and plants exposed to A. altissima VOCs (TR). The
   temporal profiles are presented as line graphs based on the two
   conditions under study: green for the WT and red for the TR. The graphs
   represent the normalised relative abundance profiles of the organic
   acids and polyamines with a good classification score and previously
   resulted significantly affected by the two-way ANOVA (p ≤ 0.05). N = 4.

Figure 7.

   [110]Figure 7
   [111]Open in a new tab

   Sugars with distinctive temporal profiles between control plants (WT)
   and plants exposed to A. altissima VOCs (TR). The temporal profiles are
   presented as line graphs based on the two conditions under study: green
   for the WT and red for the TR. The graphs represent the normalised
   relative abundance profiles of the sugars with a good classification
   score and previously resulted significantly affected by the two-way
   ANOVA (p ≤ 0.05). N = 4.

Figure 8.

   [112]Figure 8
   [113]Open in a new tab

   Miscellaneous (molecules belonging to different classes) with
   distinctive temporal profiles between control plants (WT) and plants
   exposed to A. altissima VOCs (TR). The temporal profiles are presented
   as line graphs based on the two conditions under study: green for the
   WT and red for the TR. The graphs represent the normalised relative
   abundance profiles of the different classes of compounds with a good
   classification score and previously resulted significantly affected by
   the two-way ANOVA (p ≤ 0.05). N = 4.

4. Discussion

   Rapid seed germination and establishment are vital for survival in
   natural ecosystems, providing a competitive advantage in securing
   light, water, and nutrients. Allelopathic species enhance this
   advantage by releasing chemical compounds that inhibit or delay the
   germination of nearby plants, thus suppressing potential competitors
   and enabling their dominance. In our experiments, we examined the
   effects of VOCs from the fresh leaves of A. altissima on the total
   germination and germination speed of the weed B. pilosa. As in other
   species where A. altissima leaf extracts were tested [[114]27], the
   VOCs released by A. altissima significantly reduced the total
   germination by up to 60% at the highest amount of plant material after
   eight days of treatment. Notably, concerning the speed of germination,
   even low concentrations markedly slowed germination speed, reaching 73%
   inhibition at the highest concentration. In addition, during the
   experiments we noted that the inhibition of germination observed at the
   highest concentration was eventually lost, since the seeds germinated
   after 13 days. Previous experiments already described that the duration
   of the inhibitory effect of VOCs could be lost over time.
   Camacho-Coronel et al. [[115]28] found that VOCs, like 4Z-heptenol,
   farnesene, limonene, and 2E-decenal, inhibited germination rates to
   less than 25% of control levels after 15 days, although their
   concentrations decreased over time. This suggests that, while the
   initial impact of these volatiles is strong, their long-term efficacy
   may diminish, necessitating further investigation into how these
   compounds can be effectively utilised over extended periods [[116]28].

   Looking at the VOCs characterised during A. altissima profiling,
   several compounds known for their allelopathic activity were
   identified. For example, Bradow and Connick [[117]29] reported that
   2-hexanal was the strongest volatile seed inhibitor released by plant
   residues, followed by nonanal. Even the compound 3-Hexen-1-ol, a green
   leaf volatile, plays a significant role in plant defence mechanisms and
   has been studied for its phytotoxic effects on various plant species.
   This compound is released rapidly in response to herbivore attacks,
   serving as a signal for both direct and indirect plant defences. For
   instance, Liao et al. [[118]30] demonstrated that (Z)-3-hexen-1-ol is
   emitted within minutes after herbivore damage, indicating its role as a
   pest-inducible secondary defence compound in plants like tea (Camellia
   sinensis). Similarly, the emission of (Z)-3-hexen-1-ol has been linked
   to the activation of defence pathways in other plants, including rice,
   where it enhances resistance to pests such as the brown planthopper
   [[119]31]. Besides these compounds, it should be highlighted that the
   phytotoxic potential of several VOCs reported in [120]Table 2 was
   largely discussed in the following review [[121]32].

   Given that germination involves coordinated activation of multiple
   metabolic pathways, a metabolomic time-course experiment was conducted
   to gain further insight into the possible modes of action of these
   VOCs. Cluster analysis revealed distinct metabolome evolution patterns
   between treated and untreated seeds, complementing the pathway and
   enrichment analysis that suggested metabolic alterations. In the
   untreated seeds, the metabolome developed over time, with the seeds
   spreading across three clusters corresponding to different
   developmental stages. The treated seeds, however, showed no progression
   initially, as T1 and T2 grouped with the control seeds at T0. By T3 and
   T4, the treated seeds clustered with the control seeds at later stages,
   indicating a recovery in germination. These results prompted us to
   identify the pathways affected by treatment and examine the specific
   metabolites involved.

   In this context, the interplay between energy metabolism, amino acid
   catabolism, and synthesising essential metabolites is crucial for
   successful seed germination and subsequent seedling development. Both
   enrichment analysis and pathway analysis highlighted a significant
   alteration of several pathways involved with germination. For instance,
   among the affected pathways, glycerolipid metabolism is crucial,
   providing fatty acids for energy via β-oxidation, essential for early
   embryo growth [[122]33]. Propanoate metabolism also aids energy
   production, particularly in seeds utilising fatty acid reserves
   [[123]33]. Gluconeogenesis converts stored carbohydrates into glucose,
   supplying energy in early germination [[124]34]. Central to energy
   production, the TCA cycle and glycolysis convert substrates into energy
   and essential intermediates, supporting amino acid synthesis
   [[125]34,[126]35]. Amino acid metabolism, such as valine, leucine, and
   isoleucine catabolism, supplies energy and carbon skeletons, while
   arginine and proline metabolism links nitrogen and carbon cycles via
   TCA cycle intermediates [[127]36,[128]37]. Galactose, starch, and
   sucrose metabolism provides energy and components for cell expansion
   [[129]38]. Additionally, glycine, serine, and threonine metabolism
   contributes precursors necessary for protein synthesis and other
   metabolic activities [[130]39].

   During seed germination, storage compounds, like starch, proteins, and
   lipids, are broken down to fuel growth. Starch, for example, is
   converted into simple sugars, like glucose and maltose, providing
   energy for the embryo. This breakdown is catalysed by enzymes, such as
   α-amylase, which degrades starch into glucose, a critical energy source
   and precursor for essential metabolites [[131]40,[132]41]. Disruption
   of this carbohydrate metabolism by environmental stressors or
   inhibitors can significantly delay or halt germination [[133]42], and
   in our experiments, starch and sucrose metabolism resulted among the
   most impacted pathways. In fact, in the control seeds, the amount of
   maltose and glucose significantly increased at T1 and T2 and
   successively dropped at T3 and T4. This suggests that starch was
   immediately metabolised by the seeds leading to the formation of
   maltose and glucose, which were immediately metabolised in the late
   state of the germination (T3–T4). In the treated seeds, maltose content
   reached the control content only at T2, but then, its amount was
   significantly higher at T3 and T4, suggesting a lower use of this
   compound by seed metabolism. Even the amount of glucose was slightly
   increasing compared to the control in the treated seeds, and only at T4
   did its content reach the control level, further suggesting that the
   treatment slowed down the metabolism. At T1, we observed a decrease in
   palmitic acid in the treated plants, which later returned to the
   control level by T2. This finding suggests that, with limited glucose
   as the primary substrate for glycolysis and energy production, the
   treated plants may have activated alternative pathways, such as
   β-oxidation, to generate energy. This hypothesis is further supported
   by the observed reduction in free phosphate in the treated plants at
   T2, likely used in ATP production. In contrast, in the control plants,
   this decrease was already evident at T1. Additionally, changes in
   glycerolipid and propanoate metabolism align with activating these
   alternative energy pathways.

   Besides starch and sucrose metabolism, galactose metabolism was one of
   the most affected pathways. Galactose metabolism plays a significant
   role during seed germination, particularly in the context of the
   breakdown of raffinose family oligosaccharides (RFOs), which are
   crucial for providing energy and carbon skeletons necessary for
   germination [[134]43]. RFOs, such as raffinose and stachyose, are
   hydrolysed by specific enzymes, releasing galactose and sucrose. This
   metabolic pathway is essential as it provides energy and influences the
   osmotic balance within the seed, which is critical during the initial
   stages of germination [[135]44,[136]45]. In our experiment, the content
   of raffinose decreased faster in the control seeds than in the treated
   seeds, and at the same time, the content of galactose increased in the
   control seeds at T1, probably as a result of raffinose degradation, but
   no accumulation of sucrose was observed.

   In addition to sugars, amino acids also play a pivotal role in seed
   germination, especially under stressful conditions. The degradation of
   seed-storage proteins releases amino acids, which can be utilised to
   synthesise new proteins and other metabolites essential for growth
   [[137]46,[138]47]. For instance, the GABA shunt pathway, which involves
   the conversion of glutamate to gamma-aminobutyric acid, is activated
   during germination under salt stress [[139]48]. In our experiments, we
   observed a reduction in glutamate and glutamine concentrations in the
   treated seeds compared to the control and a significant increase in
   GABA levels at T3 and T4. This suggests that plants respond to
   VOC-mediated stress by activating this pathway, which helps mitigate
   oxidative stress and supports the metabolic balance needed for
   successful germination.

   During germination, the release and levels of specific amino acids
   change significantly. Guo et al. [[140]41] observed that lysine
   initially rises and then sharply declines after 72 h, suggesting varied
   amino acid utilisation across germination stages. Similarly, Toyoizumi
   et al. [[141]49] found notable increases in amino acids, such as GABA,
   asparagine, glycine, arginine, alanine, proline, valine, isoleucine,
   leucine, lysine, and phenylalanine, while glutamate and aspartate
   levels decreased. Besides the previously observed changes induced by
   the treatment in GABA, glutamate, and glutamine content, the content of
   the other amino acids significantly changed in the treated seeds. In
   fact, in the control-developing seed, an accumulation of several amino
   acids was observed over time. As observed by Guo et al. [[142]41], the
   lysine content in our experiment initially rose from T0 to T3 before
   declining. In the treated plants, however, we observed an initial
   decrease in lysine content during early germination, followed by an
   increase at T3 and T4, mirroring the accumulation pattern observed in
   the control seeds at T1 and T2. This suggests that all the metabolic
   processes connected to lysine accumulation started later in the treated
   seeds. A similar trend was observed in other amino acids, such as
   methionine and threonine, which mirrored the control pattern but
   remained at lower levels until T3, reaching the control values only at
   T4. Amino acids, like lysine, methionine, and threonine, derived from
   aspartate, are associated with germination efficiency and subsequent
   growth [[143]50]. Aspartic acid concentration, which decreased at T1 in
   the control seeds, showed a reduction in the treated seeds only by T3,
   suggesting that this delay might result from slower aspartate
   degradation. This aligns with our findings, where pathway analysis
   indicated alanine, aspartate, and glutamate metabolism as the most
   affected pathways.

5. Conclusions

   This study reveals that Ailanthus altissima VOCs not only inhibit
   Bidens pilosa germination but also significantly slow its metabolism.
   Metabolomic analysis shows delayed starch breakdown, resulting in lower
   glucose and maltose availability for energy, as well as reduced
   glycolysis and TCA cycle activity, suggesting limited energy
   production. Amino acid metabolism, particularly involving lysine and
   threonine, also slowed, delaying critical protein synthesis needed for
   growth.

   This metabolic deceleration likely weakens B. pilosa’s competitive edge
   during early growth stages, thus enhancing A. altissima’s invasive
   advantage by suppressing nearby competitors. These findings emphasise
   the potential of VOC-mediated allelopathy as a mechanism for A.
   altissima’s ecological dominance and invite further exploration into
   VOC applications for invasive species control.

Supplementary Materials

   The following supporting information can be downloaded at:
   [144]https://www.mdpi.com/article/10.3390/metabo15010012/s1, Table S1:
   Volatilome and untargeted-metabolomics raw data and statistical
   analysis.
   [145]metabolites-15-00012-s001.zip^ (1.3MB, zip)

Author Contributions

   Conceptualisation, F.A. and L.B.; methodology, F.A. and L.B.;
   validation, D.M.M.; formal analysis, F.A., L.B. and D.M.M.;
   investigation, F.A., L.B. and D.M.M.; supervision and project
   administration, F.A. and L.B.; funding acquisition, F.A.;
   writing—original draft, F.A.; writing—review and editing, F.A., L.B.
   and D.M.M. All authors have read and agreed to the published version of
   the manuscript.

Institutional Review Board Statement

   Not applicable.

Informed Consent Statement

   Not applicable.

Data Availability Statement

   The data presented in this study are available in the article and
   [146]Supplementary Material.

Conflicts of Interest

   The authors declare no conflicts of interest.

Funding Statement

   The authors declare that this work was partially funded by the Agritech
   National Research Centre, received funding from the European Union
   NextGenerationEU (PIANO NAZIONALE DI RIPRESA E RESILIENZA
   (PNRR)–MISSIONE 4 COMPONENTE 2, INVESTIMENTO 1.4–D.D. 1032 17/06/2022,
   CN00000022) and partially by The European EU-Horizon project “AGROSUS:
   AGROecological strategies for SUStainable weed management in key
   European crop” funded the research under Grant Agreement Number
   101084084.

Footnotes

   Disclaimer/Publisher’s Note: The statements, opinions and data
   contained in all publications are solely those of the individual
   author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI
   and/or the editor(s) disclaim responsibility for any injury to people
   or property resulting from any ideas, methods, instructions or products
   referred to in the content.

References